Title: Non-destructive test method of rock bolt based on D-S evidence and spectral kurtosis

Authors: Xiaoyun Sun; Haiqing Zheng; Zhiyuan Wang; Jianpeng Bian; Hui Xing; Mingming Wang

Addresses: School of Electrical and Electronics Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043, China ' School of Electrical and Electronics Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043, China ' School of Electrical and Electronics Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043, China ' School of Electrical and Electronics Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043, China ' School of Electrical and Electronics Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043, China ' School of Electrical and Electronics Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043, China

Abstract: The length of the rock bolt is an important factor to evaluate the quality of anchor. According to the fact that the calculated value of anchor length is far different from the actual situation due to a lot of noises, a de-noising method based on Empirical Mode Decomposition (EMD) and Spectral Kurtosis (SK) is proposed in this paper, to filter noise, and improve accuracy of calculating anchor length. The fundamental idea is that calculating spectral kurtosis for each component after EMD, the larger spectral kurtosis are used to reconstruct signal to improve signal to noise ratio. The analysis results demonstrate that the method can improve spectral kurtosis of the reconstructed signal and decrease the error of anchor length. In addition, D-S evidence is introduced to realise high precision computation for anchor length by data fusion of wavelet threshold de-noising and spectral kurtosis filter.

Keywords: spectral kurtosis; anchor; EMD decomposition; wavelet de-noising; D-S data fusion.

DOI: 10.1504/IJCAT.2018.091640

International Journal of Computer Applications in Technology, 2018 Vol.57 No.2, pp.167 - 176

Available online: 30 Apr 2018 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article